agentleFS
sonichisonichi/sutando/skills/marketplace/SKILL.md

marketplace

Skill396 starsChanged 6 months agoMITView on GitHub

Install marketplace skills, activate cloud tools, check whether installed skills are up to date, or remove them — without the owner opening the Marketplace. Use when the owner asks to install/activate/enable/remove a skill or cloud tool, asks "are my skills up to date?", or hands you a use-case doc and asks you to set up what it needs.

9 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Add --json for machine-readable output.

    line 26
  2. Very specific:

    Plan: Run install <all slugs> without --yes.

    line 34
  3. Very specific:

    When the output says the desktop app hasn't connected the engine to AG2 Cloud yet (restart_after_sign_in: true), the tools need that same …

    line 58
  4. Distinctive:

    Never restart the core yourself.

    line 62
  5. Very specific:

    Run status. Skills show ok, missing (owned but not on this machine), outdated (local version behind the marketplace), disabled, or bundled …

    line 71

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Marketplace

The owner is already signed in to Sutando Cloud (desktop onboarding), so everything the Marketplace UI does can be done from here: activate a cloud tool, install a local skill, update skills, remove them.

The script lives next to this file. Run it with python3:

M="<this skill's directory>/scripts/marketplace.py"
python3 "$M" find "lead enrichment"              # search (add --kind skill|cloud_tool|connector)
python3 "$M" status                              # owned vs on this machine
python3 "$M" install intent-leads campaign-runner   # PLAN only — writes nothing
python3 "$M" install intent-leads campaign-runner --yes
python3 "$M" update                              # list missing / outdated skills
python3 "$M" update --yes                        # re-fetch them (never charges)
python3 "$M" uninstall live-preview              # PLAN only
python3 "$M" uninstall live-preview --yes

Add --json for machine-readable output. One install call takes skills and cloud tools together; the script works out which is which.

How to run a setup

  1. Collect the slugs. From the owner's request, or from a use-case doc (read it; docs list what they need under a Requires: heading). If a name isn't an exact slug, run find first.
  2. Plan. Run install <all slugs> without --yes. Show the owner the plan in a sentence or two (what gets installed or activated, what's already set up, what's skipped and why).
  3. Confirm when it costs something. Exit code 3 means the plan spends credits once, or activates a cloud tool that charges per use (the plan says "leads charges per use: 5 credits per result"): tell the owner the price and get their explicit OK first. Exit code 0 means it's free: go ahead. Never pass --yes to uninstall without the owner's OK.
  4. Apply. Re-run the same command with --yes. One failed item doesn't stop the rest.
  5. Report. Say what's ready and what failed, using the reasons the script prints (not enough credits, plan too low, not found…).
  6. Use it now; restart only when told. A newly activated cloud tool is usable at once through station_find / station_call (the output says "usable now through station_call", usable_now in JSON). Only if the output says RESTART REQUIRED (restart_required: true) did the running engine start without the Station, or for another account; then tell the owner exactly once, after everything else:

    New cloud tools are active on your account, but I can only use them after an engine restart: open Agent settings (the bot icon, bottom left), scroll down to Runtime, and click Restart engine. Restarting ends my current session; your tasks and files are kept.

    When the output says the desktop app hasn't connected the engine to AG2 Cloud yet (restart_after_sign_in: true), the tools need that same one restart once the owner signs in; re-check with status then.

    Never restart the core yourself. Installed skills need no restart; they are usable right away.

  7. Price before the first paid call. station_find marks a metered tool confirm_before_call: true. Before the first station_call to such a tool in a conversation, state its price from pricing ("5 credits per result") and wait for the owner's OK.

"Are my skills up to date?"

Run status. Skills show ok, missing (owned but not on this machine), outdated (local version behind the marketplace), disabled, or bundled (ships with Sutando). If anything is missing or outdated, offer update --yes. Updating never charges.

Things to know

  • Connectors (Gmail, Google Calendar, Slack, …) are not installed here: install skips them with a note, and find / install match them by exact slug. Connect them from chat with the connect-apps skill, which sends the owner a Connect card and picks the request up once they sign in.
  • Paid skills charge once. Uninstalling one and reinstalling it charges again.
  • Skills install as real directories under $CLAUDE_CONFIG_DIR/skills/<slug>/, verified against the marketplace's sha256. Unsigned bundles are refused. Skills bundled with Sutando are never replaced.
  • Errors:
    • "not signed in" / "session expired": the owner signs in again from the desktop app.
    • Exit 2: a setup problem to relay, not something to retry.

Writing a use-case doc that Sutando can set up

List the slugs under a Requires: heading:

## Requires
- Cloud tools: icp-mapper, intent-leads, lead-enricher, email-drafter
- Skills: campaign-runner, campaign-responder

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.